How to edit a podcast video, from raw recording to clips

How do you edit a podcast video, step by step?

Video podcast editing is two edits in one: a sound edit that has to feel like a natural conversation, and a picture edit laid over it. In this order, each step makes the next easier:

Sync everything to one timeline: every camera, microphone and recorder.
Cut on the sound. Take out false starts, long pauses, tangents and talk-over, listening rather than watching.
Choose the angles. The person speaking, a reaction where it says something, a wide shot when everyone talks at once.
Clean and level each voice so nobody is louder than anybody else.
Set the loudness for where the episode is going.
Add chapters, titles and captions.
Cut the clips for shorts and social.

Cutting picture before sound is the classic mistake: every later sound fix then breaks an angle change.

How do you sync several cameras and microphones?

By their sound. Every camera records a scratch track that software can line up with the good audio from the microphones, and one clap in front of every camera before you start gives them a sharp moment to agree on.

A man in a beanie claps his hands once between two microphones in a basement studio, two cameras on tripods facing him and his co-host laughing behind. — One clap before recording gives every camera and microphone the same sharp sound to line up on.

Then watch for drift. No two devices keep exactly the same time: clocks 100 parts per million apart drift 0.36 seconds over an hour, about eleven frames, enough to put lips visibly out of step by the end of an episode.

Sharpy syncs by sound with Synchronize by Sound in the window, or by timecode with Clip ▸ Synchronize by Timecode when camera and recorder both carry it. It measures the drift between two devices' clocks and takes it out, and it aligns a second camera by what both microphones heard, so mouths stay on their words. Two 21-minute recordings sync in 7.5 seconds on the base Mac mini M4. How to sync audio and video covers the craft in full.

How do you edit a multicam podcast?

Cut the conversation first, then switch angles over it. Multicam podcast editing feels natural when you:

Show who is speaking, cutting to them just after they start rather than before.
Use reactions on purpose. A laugh or a raised eyebrow earns a cut; a blank stare does not.
Go wide when people talk over each other, so viewers see everyone at once.
Switch less than you think. Constant changes tire the eye over an hour.

In Sharpy, a multicam clip carries every angle, with picture and sound angles kept apart, and you or the agent switch angles after the cut. Sharpy asks how many people are speaking, turns each speaker's turns into angle cuts or crops around each speaker, and adds a lower third at each speaker's first turn. Tell it the number of speakers and the count is exact. Measured on the base Mac mini M4 against VoxConverse, a public set of 216 recordings with 1 to 20 speakers, its split by speaker is wrong 3.67% of the time. In a split-screen layout, each panel freezes while another person talks and comes back on its next turn.

How do you cut crosstalk, fillers and dead air without losing the rhythm?

Cut what gets in the way of the conversation and keep what makes it one. The laughs, quick agreements and short reactions stay; the ums that break a sentence, the restarts and the long silences while someone thinks go. Crosstalk is a judgement call: keep it when it is lively and both voices are clear, cut it back to one voice when nobody can follow either. Split edits smooth the joins: let the next voice start under the end of the last shot (a J cut), or carry on over the listener (an L cut).

Sharpy puts everything an editor would look at into one review list, each item with its reason and Hear, Cut or Shorten, and Keep:

What it finds	What it does
Fillers: um, uh, erm	Cuts them; "mm-hmm" and "uh-huh" mean yes and stay
Filler phrases: you know, I mean, sort of	Lists them for you to check
Repeated words and restarts	Keeps the last, complete attempt
A line said twice	Keeps one, and you choose which
Long gaps	Shortens them to the recording's quickest pause at that kind of place in a sentence, never to nothing

Restarts and retakes are found per speaker, including where one person echoes another, and pauses the story needs, before a reveal or a change of subject, are left off the list. A gap a cut leaves on a voice track is filled with the room's own tone rather than digital silence. Removing filler words and dead air explains the rules.

How do you level each voice and set podcast loudness?

Level the voices first, then the whole mix. Guests sit at different distances from their microphones, so each voice needs its own gain before anything else.

Sharpy levels dialogue clip by clip to −23 LUFS, raising quiet speech without lifting the room in the pauses, and reports what it did, for example "speech at −34.6 LUFS, +11.6 dB to −23 LUFS". Two lines recorded 17.5 dB apart both came out at −23.0. Send every voice to one bus in the mixer and one compressor holds them together; voice clean-up takes out room noise. Then set the episode's loudness for where it goes:

Where it goes	Loudness	Whose figure
Apple Podcasts	−16 LKFS ±1 dB, true peak no higher than −1 dBFS	Apple Podcasts for Creators
YouTube and streaming video	−14 LUFS	Sharpy's streaming destinations
Broadcast	−23 LUFS	EBU R 128

LKFS and LUFS are the same unit. Sharpy measures the finished file, not a prediction, and lands on the target you name: a minute of its maker's own recording came out at −14.07 LUFS with a true peak of −1.14 dBTP. What LUFS for YouTube explains why −14.

How do you add chapters and cut clips from the episode?

Mark them as you review. A chapter at each change of topic helps viewers jump around; YouTube shows chapters when the first starts at 00:00, there are at least three, and each lasts 10 seconds or more.

In Sharpy, a chapter is a kind of marker, and chapters can also be made from the transcript. They go into the file as a chapter list players read, and into the description as YouTube timestamps; where an edit breaks YouTube's rules, Sharpy says so instead.

For clips, lay a range marker over each moment worth sharing and choose File ▸ Export Each Marked Stretch…: each comes out as its own file, named after its marker, with its captions beside it. For vertical shorts, see how to turn a long video into shorts.

Frequently asked questions

What is the best editing tool for podcasts?

One that edits sound and picture together, syncs several cameras by their sound and lets you edit by the transcript. Sharpy does all three on your Mac and makes the first cut itself, with every decision listed; Descript, Premiere Pro, DaVinci Resolve and Final Cut Pro are the other common choices.

Can I use AI to edit a podcast?

Yes. AI is good at the slow parts: transcribing, splitting speakers, finding fillers, repeats and long pauses, and suggesting clips. Sharpy does these on your Mac and gives every decision a reason, so you check its work instead of redoing it.

How long does it take to edit a video podcast?

By hand, often several hours for a one-hour episode. Sharpy reads a recording 1.7 times faster than real time on the base Mac mini M4, every word, pause and shot, so it has been through a one-hour episode in about 35 minutes.

Can Sharpy handle a long episode?

Yes. It transcribes and splits by speaker recordings up to four hours long on a 16 GB Mac, with nothing uploaded and nothing metered.

Do I need a separate audio recorder?

Not if your microphones go into a good interface, but a recorder makes a safety copy of every voice. Sharpy lines a recorder up with the cameras by sound or by timecode either way.

Sources

Checked on 18 October 2026.